Left: | ||
Right: |
LEFT | RIGHT |
---|---|
1 /* | 1 /* |
2 * This file is part of Adblock Plus <https://adblockplus.org/>, | 2 * This file is part of Adblock Plus <https://adblockplus.org/>, |
3 * Copyright (C) 2006-present eyeo GmbH | 3 * Copyright (C) 2006-present eyeo GmbH |
4 * | 4 * |
5 * Adblock Plus is free software: you can redistribute it and/or modify | 5 * Adblock Plus is free software: you can redistribute it and/or modify |
6 * it under the terms of the GNU General Public License version 3 as | 6 * it under the terms of the GNU General Public License version 3 as |
7 * published by the Free Software Foundation. | 7 * published by the Free Software Foundation. |
8 * | 8 * |
9 * Adblock Plus is distributed in the hope that it will be useful, | 9 * Adblock Plus is distributed in the hope that it will be useful, |
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of | 10 * but WITHOUT ANY WARRANTY; without even the implied warranty of |
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | 11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
12 * GNU General Public License for more details. | 12 * GNU General Public License for more details. |
13 * | 13 * |
14 * You should have received a copy of the GNU General Public License | 14 * You should have received a copy of the GNU General Public License |
15 * along with Adblock Plus. If not, see <http://www.gnu.org/licenses/>. | 15 * along with Adblock Plus. If not, see <http://www.gnu.org/licenses/>. |
16 */ | 16 */ |
17 | 17 |
18 "use strict"; | 18 "use strict"; |
19 | 19 |
20 /** | 20 /** |
21 * @fileOverview Matcher class implementing matching addresses against | 21 * @fileOverview Matcher class implementing matching addresses against |
22 * a list of filters. | 22 * a list of filters. |
23 */ | 23 */ |
24 | 24 |
25 const {Filter, WhitelistFilter} = require("./filterClasses"); | 25 const {WhitelistFilter} = require("./filterClasses"); |
26 | 26 |
27 /** | |
28 * Blacklist/whitelist filter matching | |
29 */ | |
27 class Matcher | 30 class Matcher |
28 { | 31 { |
29 /** | |
Manish Jethani
2018/08/31 13:01:34
I think this rather sounds like the description of
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
30 * Blacklist/whitelist filter matching | |
31 */ | |
32 constructor() | 32 constructor() |
33 { | 33 { |
34 /** | 34 /** |
35 * Lookup table for filters by their associated keyword | 35 * Lookup table for filters by their associated keyword |
36 * @type {Map.<string,(Filter|Filter[])>} | 36 * @type {Map.<string,(Filter|Filter[])>} |
37 */ | 37 */ |
38 this.filterByKeyword = null; | 38 this.filterByKeyword = new Map(); |
39 | 39 |
40 /** | 40 /** |
41 * Lookup table for keywords by the filter | 41 * Lookup table for keywords by the filter |
42 * @type {Map.<Filter,string>} | 42 * @type {Map.<Filter,string>} |
43 */ | 43 */ |
44 this.keywordByFilter = null; | 44 this.keywordByFilter = new Map(); |
45 this.clear(); | |
Manish Jethani
2018/08/31 13:01:35
Let's leave a line before this.clear().
Actually
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
46 } | 45 } |
47 | 46 |
48 /** | 47 /** |
49 * Removes all known filters | 48 * Removes all known filters |
50 */ | 49 */ |
51 clear() | 50 clear() |
52 { | 51 { |
53 this.filterByKeyword = new Map(); | 52 this.filterByKeyword.clear(); |
Manish Jethani
2018/08/31 13:01:35
Since these are Map objects now, we can just call
Jon Sonesen
2018/09/02 16:43:19
Done.
| |
54 this.keywordByFilter = new Map(); | 53 this.keywordByFilter.clear(); |
55 } | 54 } |
56 | 55 |
57 /** | 56 /** |
58 * Adds a filter to the matcher | 57 * Adds a filter to the matcher |
59 * @param {RegExpFilter} filter | 58 * @param {RegExpFilter} filter |
60 */ | 59 */ |
61 add(filter) | 60 add(filter) |
62 { | 61 { |
63 if (this.keywordByFilter.has(filter)) | 62 if (this.keywordByFilter.has(filter)) |
64 return; | 63 return; |
(...skipping 33 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
98 this.filterByKeyword.set(keyword, list[0]); | 97 this.filterByKeyword.set(keyword, list[0]); |
99 } | 98 } |
100 } | 99 } |
101 | 100 |
102 this.keywordByFilter.delete(filter); | 101 this.keywordByFilter.delete(filter); |
103 } | 102 } |
104 | 103 |
105 /** | 104 /** |
106 * Chooses a keyword to be associated with the filter | 105 * Chooses a keyword to be associated with the filter |
107 * @param {Filter} filter | 106 * @param {Filter} filter |
108 * @return {string} keyword or an empty string if no keyword could be found | 107 * @returns {string} keyword or an empty string if no keyword could be found |
Manish Jethani
2018/08/31 13:01:34
Let's s/@return/@returns/ here and throughout the
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
109 */ | 108 */ |
110 findKeyword(filter) | 109 findKeyword(filter) |
111 { | 110 { |
112 let result = ""; | 111 let result = ""; |
113 let {text} = filter; | 112 let {pattern} = filter; |
114 if (Filter.regexpRegExp.test(text)) | 113 if (pattern == null) |
115 return result; | 114 return result; |
116 | 115 |
117 // Remove options | 116 let candidates = pattern.toLowerCase().match( |
118 let match = Filter.optionsRegExp.exec(text); | |
119 if (match) | |
120 text = match.input.substr(0, match.index); | |
121 | |
122 // Remove whitelist marker | |
123 if (text[0] == "@" && text[1] == "@") | |
124 text = text.substr(2); | |
125 | |
126 let candidates = text.toLowerCase().match( | |
127 /[^a-z0-9%*][a-z0-9%]{3,}(?=[^a-z0-9%*])/g | 117 /[^a-z0-9%*][a-z0-9%]{3,}(?=[^a-z0-9%*])/g |
128 ); | 118 ); |
129 if (!candidates) | 119 if (!candidates) |
130 return result; | 120 return result; |
131 | 121 |
132 let hash = this.filterByKeyword; | 122 let hash = this.filterByKeyword; |
133 let resultCount = 0xFFFFFF; | 123 let resultCount = 0xFFFFFF; |
134 let resultLength = 0; | 124 let resultLength = 0; |
135 for (let i = 0, l = candidates.length; i < l; i++) | 125 for (let i = 0, l = candidates.length; i < l; i++) |
136 { | 126 { |
137 let candidate = candidates[i].substr(1); | 127 let candidate = candidates[i].substr(1); |
138 let filters = hash.get(candidate); | 128 let filters = hash.get(candidate); |
139 let count = typeof filters != "undefined" ? filters.length : 0; | 129 let count = typeof filters != "undefined" ? filters.length : 0; |
140 if (count < resultCount || | 130 if (count < resultCount || |
141 (count == resultCount && candidate.length > resultLength)) | 131 (count == resultCount && candidate.length > resultLength)) |
142 { | 132 { |
143 result = candidate; | 133 result = candidate; |
144 resultCount = count; | 134 resultCount = count; |
145 resultLength = candidate.length; | 135 resultLength = candidate.length; |
146 } | 136 } |
147 } | 137 } |
148 return result; | 138 return result; |
149 } | 139 } |
150 | 140 |
151 /** | 141 /** |
152 * Checks whether a particular filter is being matched against. | 142 * Checks whether a particular filter is being matched against. |
153 * @param {RegExpFilter} filter | 143 * @param {RegExpFilter} filter |
154 * @return {boolean} | 144 * @returns {boolean} |
155 */ | 145 */ |
156 hasFilter(filter) | 146 hasFilter(filter) |
157 { | 147 { |
158 return this.keywordByFilter.has(filter); | 148 return this.keywordByFilter.has(filter); |
159 } | 149 } |
160 | 150 |
161 /** | 151 /** |
162 * Returns the keyword used for a filter, null for unknown filters. | 152 * Returns the keyword used for a filter, <code>null</code> |
153 * for unknown filters. | |
163 * @param {RegExpFilter} filter | 154 * @param {RegExpFilter} filter |
164 * @return {?string} | 155 * @returns {?string} |
165 */ | 156 */ |
166 getKeywordForFilter(filter) | 157 getKeywordForFilter(filter) |
167 { | 158 { |
168 let keyword = this.keywordByFilter.get(filter); | 159 let keyword = this.keywordByFilter.get(filter); |
169 return typeof keyword != "undefined" ? keyword : null; | 160 return typeof keyword != "undefined" ? keyword : null; |
170 } | 161 } |
171 | 162 |
172 /** | 163 /** |
173 * Checks whether the entries for a particular keyword match a URL | 164 * Checks whether the entries for a particular keyword match a URL |
174 * @param {string} keyword | 165 * @param {string} keyword |
175 * @param {string} location | 166 * @param {string} location |
176 * @param {number} typeMask | 167 * @param {number} typeMask |
177 * @param {string} docDomain | 168 * @param {string} [docDomain] |
Manish Jethani
2018/08/31 13:01:34
docDomain, thirdParty, sitekey, and specificOnly a
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
178 * @param {boolean} thirdParty | 169 * @param {boolean} [thirdParty] |
179 * @param {string} sitekey | 170 * @param {string} [sitekey] |
180 * @param {boolean} specificOnly | 171 * @param {boolean} [specificOnly] |
181 * @return {?Filter} | 172 * @returns {?Filter} |
182 */ | 173 */ |
183 _checkEntryMatch(keyword, location, typeMask, docDomain, thirdParty, sitekey, | 174 _checkEntryMatch(keyword, location, typeMask, docDomain, thirdParty, sitekey, |
184 specificOnly) | 175 specificOnly) |
185 { | 176 { |
186 let list = this.filterByKeyword.get(keyword); | 177 let list = this.filterByKeyword.get(keyword); |
187 if (typeof list == "undefined") | 178 if (typeof list == "undefined") |
188 return null; | 179 return null; |
189 for (let i = 0; i < list.length; i++) | 180 for (let i = 0; i < list.length; i++) |
190 { | 181 { |
191 let filter = list[i]; | 182 let filter = list[i]; |
192 | 183 |
193 if (specificOnly && filter.isGeneric() && | 184 if (specificOnly && filter.isGeneric() && |
194 !(filter instanceof WhitelistFilter)) | 185 !(filter instanceof WhitelistFilter)) |
195 continue; | 186 continue; |
196 | 187 |
197 if (filter.matches(location, typeMask, docDomain, thirdParty, sitekey)) | 188 if (filter.matches(location, typeMask, docDomain, thirdParty, sitekey)) |
198 return filter; | 189 return filter; |
199 } | 190 } |
200 return null; | 191 return null; |
201 } | 192 } |
202 | 193 |
203 /** | 194 /** |
204 * Tests whether the URL matches any of the known filters | 195 * Tests whether the URL matches any of the known filters |
205 * @param {string} location | 196 * @param {string} location |
206 * URL to be tested | 197 * URL to be tested |
207 * @param {number} typeMask | 198 * @param {number} typeMask |
208 * bitmask of content / request types to match | 199 * bitmask of content / request types to match |
209 * @param {string} docDomain | 200 * @param {string} [docDomain] |
Manish Jethani
2018/08/31 13:01:35
See comment above about optional parameters.
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
210 * domain name of the document that loads the URL | 201 * domain name of the document that loads the URL |
211 * @param {boolean} thirdParty | 202 * @param {boolean} [thirdParty] |
212 * should be true if the URL is a third-party request | 203 * should be true if the URL is a third-party request |
213 * @param {string} sitekey | 204 * @param {string} [sitekey] |
214 * public key provided by the document | 205 * public key provided by the document |
215 * @param {boolean} specificOnly | 206 * @param {boolean} [specificOnly] |
216 * should be true if generic matches should be ignored | 207 * should be <code>true</code> if generic matches should be ignored |
Manish Jethani
2018/08/31 13:01:35
Let's make this <code>true</code>.
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
217 * @return {?RegExpFilter} | 208 * @returns {?RegExpFilter} |
218 * matching filter or null | 209 * matching filter or <code>null</code> |
Manish Jethani
2018/08/31 13:01:34
Let's make this <code>null</code>.
Jon Sonesen
2018/09/02 16:43:19
Done.
| |
219 */ | 210 */ |
220 matchesAny(location, typeMask, docDomain, thirdParty, sitekey, specificOnly) | 211 matchesAny(location, typeMask, docDomain, thirdParty, sitekey, specificOnly) |
221 { | 212 { |
222 let candidates = location.toLowerCase().match(/[a-z0-9%]{3,}/g); | 213 let candidates = location.toLowerCase().match(/[a-z0-9%]{3,}/g); |
223 if (candidates === null) | 214 if (candidates === null) |
224 candidates = []; | 215 candidates = []; |
225 candidates.push(""); | 216 candidates.push(""); |
226 for (let i = 0, l = candidates.length; i < l; i++) | 217 for (let i = 0, l = candidates.length; i < l; i++) |
227 { | 218 { |
228 let result = this._checkEntryMatch(candidates[i], location, typeMask, | 219 let result = this._checkEntryMatch(candidates[i], location, typeMask, |
229 docDomain, thirdParty, sitekey, | 220 docDomain, thirdParty, sitekey, |
230 specificOnly); | 221 specificOnly); |
231 if (result) | 222 if (result) |
232 return result; | 223 return result; |
233 } | 224 } |
234 | 225 |
235 return null; | 226 return null; |
236 } | 227 } |
237 } | 228 } |
Manish Jethani
2018/08/31 13:01:34
Let's leave a blank line after the class definitio
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
229 | |
238 exports.Matcher = Matcher; | 230 exports.Matcher = Matcher; |
239 | 231 |
240 | 232 |
241 | 233 /** |
234 * Combines a matcher for blocking and exception rules, automatically sorts | |
235 * rules into two {@link Matcher} instances. | |
236 */ | |
242 class CombinedMatcher | 237 class CombinedMatcher |
243 { | 238 { |
244 /** | |
245 * Combines a matcher for blocking and exception rules, automatically sorts | |
Manish Jethani
2018/08/31 13:01:34
Let's move this to above the class definition.
| |
246 * rules into two Matcher instances. | |
Manish Jethani
2018/08/31 13:01:35
Let's change `Matcher` here to `{@link Matcher}` s
| |
247 * @augments Matcher | |
Manish Jethani
2018/08/31 13:01:34
Let's remove @arguments here, it's outdated.
| |
248 */ | |
249 constructor() | 239 constructor() |
250 { | 240 { |
251 /** | 241 /** |
252 * Maximal number of matching cache entries to be kept | 242 * Maximal number of matching cache entries to be kept |
253 * @type {number} | 243 * @type {number} |
254 */ | 244 */ |
255 this.maxCacheEntries = 1000; | 245 this.maxCacheEntries = 1000; |
256 | 246 |
257 /** | 247 /** |
258 * Matcher for blocking rules. | 248 * Matcher for blocking rules. |
259 * @type {Matcher} | 249 * @type {Matcher} |
260 */ | 250 */ |
261 this.blacklist = new Matcher(); | 251 this.blacklist = new Matcher(); |
262 | 252 |
263 /** | 253 /** |
264 * Matcher for exception rules. | 254 * Matcher for exception rules. |
265 * @type {Matcher} | 255 * @type {Matcher} |
266 */ | 256 */ |
267 this.whitelist = new Matcher(); | 257 this.whitelist = new Matcher(); |
268 | 258 |
269 /** | 259 /** |
270 * Lookup table of previous matchesAny results | 260 * Lookup table of previous {@link Matcher#matchesAny} results |
Manish Jethani
2018/08/31 13:01:35
Let's make this `{@link Matcher#matchesAny}`
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
271 * @type {Map.<string,Filter>} | 261 * @type {Map.<string,Filter>} |
272 */ | 262 */ |
273 this.resultCache = new Map(); | 263 this.resultCache = new Map(); |
274 } | 264 } |
275 | 265 |
276 /** | 266 /** |
277 * @see Matcher#clear | 267 * @see Matcher#clear |
278 */ | 268 */ |
279 clear() | 269 clear() |
280 { | 270 { |
(...skipping 26 matching lines...) Expand all Loading... | |
307 this.whitelist.remove(filter); | 297 this.whitelist.remove(filter); |
308 else | 298 else |
309 this.blacklist.remove(filter); | 299 this.blacklist.remove(filter); |
310 | 300 |
311 this.resultCache.clear(); | 301 this.resultCache.clear(); |
312 } | 302 } |
313 | 303 |
314 /** | 304 /** |
315 * @see Matcher#findKeyword | 305 * @see Matcher#findKeyword |
316 * @param {Filter} filter | 306 * @param {Filter} filter |
317 * @return {string} keyword | 307 * @returns {string} keyword |
318 */ | 308 */ |
319 findKeyword(filter) | 309 findKeyword(filter) |
320 { | 310 { |
321 if (filter instanceof WhitelistFilter) | 311 if (filter instanceof WhitelistFilter) |
322 return this.whitelist.findKeyword(filter); | 312 return this.whitelist.findKeyword(filter); |
323 return this.blacklist.findKeyword(filter); | 313 return this.blacklist.findKeyword(filter); |
324 } | 314 } |
325 | 315 |
326 /** | 316 /** |
327 * @see Matcher#hasFilter | 317 * @see Matcher#hasFilter |
328 * @param {Filter} filter | 318 * @param {Filter} filter |
329 * @return {boolean} | 319 * @returns {boolean} |
330 */ | 320 */ |
331 hasFilter(filter) | 321 hasFilter(filter) |
332 { | 322 { |
333 if (filter instanceof WhitelistFilter) | 323 if (filter instanceof WhitelistFilter) |
334 return this.whitelist.hasFilter(filter); | 324 return this.whitelist.hasFilter(filter); |
335 return this.blacklist.hasFilter(filter); | 325 return this.blacklist.hasFilter(filter); |
336 } | 326 } |
337 | 327 |
338 /** | 328 /** |
339 * @see Matcher#getKeywordForFilter | 329 * @see Matcher#getKeywordForFilter |
340 * @param {Filter} filter | 330 * @param {Filter} filter |
341 * @return {string} keyword | 331 * @returns {string} keyword |
342 */ | 332 */ |
343 getKeywordForFilter(filter) | 333 getKeywordForFilter(filter) |
344 { | 334 { |
345 if (filter instanceof WhitelistFilter) | 335 if (filter instanceof WhitelistFilter) |
346 return this.whitelist.getKeywordForFilter(filter); | 336 return this.whitelist.getKeywordForFilter(filter); |
347 return this.blacklist.getKeywordForFilter(filter); | 337 return this.blacklist.getKeywordForFilter(filter); |
348 } | 338 } |
349 | 339 |
350 /** | 340 /** |
351 * Checks whether a particular filter is slow | 341 * Checks whether a particular filter is slow |
352 * @param {RegExpFilter} filter | 342 * @param {RegExpFilter} filter |
353 * @return {boolean} | 343 * @returns {boolean} |
354 */ | 344 */ |
355 isSlowFilter(filter) | 345 isSlowFilter(filter) |
356 { | 346 { |
357 let matcher = ( | 347 let matcher = ( |
358 filter instanceof WhitelistFilter ? this.whitelist : this.blacklist | 348 filter instanceof WhitelistFilter ? this.whitelist : this.blacklist |
359 ); | 349 ); |
360 if (matcher.hasFilter(filter)) | 350 if (matcher.hasFilter(filter)) |
361 return !matcher.getKeywordForFilter(filter); | 351 return !matcher.getKeywordForFilter(filter); |
362 return !matcher.findKeyword(filter); | 352 return !matcher.findKeyword(filter); |
363 } | 353 } |
(...skipping 49 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
413 result = this.matchesAnyInternal(location, typeMask, docDomain, | 403 result = this.matchesAnyInternal(location, typeMask, docDomain, |
414 thirdParty, sitekey, specificOnly); | 404 thirdParty, sitekey, specificOnly); |
415 | 405 |
416 if (this.resultCache.size >= CombinedMatcher.maxCacheEntries) | 406 if (this.resultCache.size >= CombinedMatcher.maxCacheEntries) |
417 this.resultCache.clear(); | 407 this.resultCache.clear(); |
418 | 408 |
419 this.resultCache.set(key, result); | 409 this.resultCache.set(key, result); |
420 | 410 |
421 return result; | 411 return result; |
422 } | 412 } |
423 } | 413 } |
Manish Jethani
2018/08/31 13:01:34
Let's leave a blank line here.
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
414 | |
424 exports.CombinedMatcher = CombinedMatcher; | 415 exports.CombinedMatcher = CombinedMatcher; |
425 | 416 |
426 /** | 417 /** |
427 * Shared CombinedMatcher instance that should usually be used. | 418 * Shared {@link CombinedMatcher} instance that should usually be used. |
Manish Jethani
2018/08/31 13:01:35
Let's make this `{@link CombinedMatcher}`
Jon Sonesen
2018/09/02 16:43:20
Done.
| |
428 * @type {CombinedMatcher} | 419 * @type {CombinedMatcher} |
429 */ | 420 */ |
430 exports.defaultMatcher = new CombinedMatcher(); | 421 let defaultMatcher = new CombinedMatcher(); |
422 | |
423 exports.defaultMatcher = defaultMatcher; | |
LEFT | RIGHT |